JPH021078A - Treatment of radiation picture and device for treating radiation picture - Google Patents

Abstract

PURPOSE:To improve economical and speedy diagnostic performance by performing an operation by using a prescribed function as an emphasizing coefficient and using a signal obtained by simply adding and averaging original picture signals at respective scanning points in a prescribed area as a non-sharp mask signal. CONSTITUTION:A picture element input signal SIN is sent from a gate 11, to which the signal SIN is inputted, to a memory 12 having prescribed capacity, and an arithmetic circuit 13 performs the operation based on the stored memory value. The operation result is passed through the memory 12 and outputted from the gate 11 as a picture element output signal SOUT. The respective operations are executed by a control circuit 14. Further, in an operation S'=Sorg+beta(Sorg-Sus) for obtaining a reproduced picture signal S', a non-sharp mask signal Sus is obtained by simply averaging original picture signals Sorg at the respective scanning points in an rectangular area (broad-line area) surrounded with respective two pairs of sides respectively parallel in main and sub scanning directions, and next, the signal S' is obtained by performing the operation by using the function either to simply increase according to the increase of either the value of the signal Sorg or that of the signal Sus or to simply increase and simply decrease after reading the maximum value as an emphasizing coefficient beta. Thus, the high-speed performance can be obtained, and the cost of the title device can be lowered.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an image processing method in a radiation image information recording and reproducing system used for medical diagnosis.
It's called a "fluorescent material." ) to record radiation image information thereon, and then read and reproduce this radiation image information, and record this as a final image on a recording material. It is something.

The radiation transmitted through the subject is absorbed by a phosphor to record radiation image information, which is then excited by scanning with a laser beam or the like, the emitted light is read by a photodetector, and the read radiation image information is recorded. A radiation image information recording and reproducing system is known that modulates a light beam to record a radiation image on a recording material such as a photographic film. (U.S. Patent No. 3.
(No. 859.527) This radiation image information recording and reproducing system using this phosphor is extremely superior in that it can record images over a wide radiation exposure range compared to conventional silver halide photography systems. It has high utility value, especially as an X-ray photography system for the human body.

On the other hand, X-rays are harmful to the human body when exposed to large amounts of radiation, so it goes without saying that it is desirable to obtain as much information as possible in just one X-ray photograph, but current X-ray film is not suitable for photographing. It is necessary to have both aptitude for observation and image interpretation, and since the film is designed to satisfy each of these to a certain degree, it cannot be said that the X-ray exposure range is sufficiently wide in terms of suitability for photography, and current X-ray photographic film Regarding the suitability for observation and interpretation of images, there was a problem in that the image quality was not necessarily sufficient for diagnosis.

Furthermore, the radiation image information recording and reproducing system using a phosphor disclosed in the aforementioned U.S. Pat. The points were not resolved.

In view of the above circumstances, the inventor of the present invention has proposed that in a radiation image recording method using a phosphor, a non-sharp mask process is applied when reading the radiation image information recorded on the phosphor and reproducing it on a recording material. A radiation image processing method for improving the diagnostic performance of radiation images was proposed in Japanese Patent Application Laid-Open No. 55-88740.

This method uses extremely low frequencies (hereinafter referred to as "
"Very low frequency". ) region, and that attempting to improve sharpness by emphasizing high-frequency components only emphasizes noise components in the case of radiation image processing, which actually degrades diagnostic performance. Based on the knowledge that the proportion of noise in the high frequency range is high, and that reducing the emphasis in this high frequency range makes the noise less noticeable and easier to see, we have emphasized the very low frequency components and at the same time The goal is to relatively reduce high-frequency components with large amounts of light, thereby making it possible to obtain images that are easy to see visually. After reading the radiation image information and converting it into an electrical signal, when reproducing it on the recording material, an unsharp mask signal Sus corresponding to the ultra-low frequency is determined at each scanning point and read out from the phosphor. When the Sorgs emphasis coefficient is β for the original image signal and S′ is the reproduced image signal, the signal is converted by the calculation S'-3org+β(Sorg-5us) to emphasize the frequency components above the ultra-low frequency. This is a radiation image processing method characterized by the following. Here, the ultra-low spatial frequency means a spatial frequency of about 0.5 cycles/mm or less.

Here, the unsharp mask signal Sus corresponding to the very low frequency
refers to the signal at each scanning point of an unsharp image (hereinafter referred to as "unsharp mask") that blurs the original image so that it only contains frequency components lower than ultra-low frequency components,
This non-sharp mask has a modulation transfer function of 0. A mask having a spatial frequency of 0.5 or more at a spatial frequency of 01 cycles/mtn and 0.5 or less at a spatial frequency of 0.5 cycles/m is used, and it is also difficult to create a non-sharp mask. The method is as follows: (1) The original image signal at each scanning point is stored, read out along with peripheral data according to the size of the non-sharp mask, and the average value (simple average or average value using various weighted averages) is calculated. ) (In this method, there are cases where the analog signal is created as it is, and
There are cases where the analog signal is created after A/D conversion and is created as a digital signal, and it also includes cases where the analog signal is de-sharpened with a low-pass filter only in the main scanning direction before A/D conversion, and digital signal processing is performed in the sub-scanning direction. It will be done. ), (a) If the accumulated image still remains after reading out the original image signal with a small-sized diameter light beam, etc., each scanning point is read out using a large-sized diameter light beam that matches the size of the non-sharp mask. (3) A method that takes advantage of the fact that the readout light beam gradually expands in diameter due to scattering in the phosphor layer. A method of creating an original image signal S org using the light emission signal and creating a non-sharp mask signal Sus using light emission on the side through which the light beam passes (
In this case, the size of the unsharp mask can be controlled by changing the degree of light scattering of the phosphor layer and by changing the size of the aperture that receives this light. ) can be used.

As a result of comparative studies of the non-sharp mask creation methods described in (1) to (3) above, the present inventors found that
We have found that method (1) is the most preferable in order to have Ru.

Here, l and j are the coordinates of a circular area centered on each scanning point (the number of pixels in the area is N in the diametrical direction), and aIj is a weighting coefficient that is equal in all directions. It is preferable to have a square and smooth change, and Σ aij = 1 flaw, +60.

However, if such an operation were to be performed simply, it would be necessary to perform approximately N 2 multiplications and N 2 additions for each scan point, and if N is large, the operation would be time-consuming. The disadvantage is that it is expensive and impractical. In fact, when reading out a normal radiographic image by scanning a phosphor, it is necessary to avoid losing the frequency components of the image, so although there are some differences depending on the image, it is usually Sampling rate of ~20 pixels/# (200~50μ in pixel size)
Since the non-sharp mask in Rikiki's invention corresponds to very low frequencies, it is necessary to perform calculations using a very large number of pixels to create this mask.

For example, in the case of a mask with Gaussian distributed weighting coefficients, if the pixel size is 100μ x 100μ, then f, -0
, 1 cycle/#, N is approximately 50, and fc
In the case of -0.02 cycles/rttrtt, N is approximately 250, so the calculation time becomes enormous. (Here, f means the value at which the modulation transfer function of the non-sharp mask is 0.5.) Also, averaging the circular area means changing the addition range for each scanning line. However, having to make such a judgment when performing calculations significantly complicates the calculation mechanism and is uneconomical.

An object of the present invention is to provide a radiation image processing method and apparatus that can improve diagnostic performance economically and at high speed.

In order to achieve this object, the present inventor has conducted extensive research and found that in the above image processing method, as a method for obtaining a non-sharp mask signal, two sides parallel to the main scanning direction and one in the sub-scanning direction are used. The object is to obtain a non-sharp mask signal Sus for an extremely low spatial frequency at each scanning point by simply averaging the original image signal S org at each scanning point within a rectangular area surrounded by two parallel sides. It was found that it is in line with In other words, the method of creating such a non-sharp mask has a uniform weight in a rectangular area, and therefore, compared to a mask with a Gaussian distribution weight whose weight decays smoothly, the transfer characteristic may oscillate. Despite having drawbacks such as the degree of unsharpness varying depending on the direction,
In terms of improvement in diagnostic performance, there is no real difference from the ideal mask calculation described above, and since it is a simple averaging over a rectangular area, the calculation time can be greatly reduced and the equipment size can be reduced, as will be explained later. It was discovered that significant cost reductions could be achieved.

The present invention scans a stimulable phosphor material, reads out radiation image information recorded in the phosphor material, converts it into an electrical signal, and then reproduces it as a visible image at each scanning point. The unsharp mask signal Su corresponding to the very low spatial frequency at
s is calculated, and when the original image signal read out from the phosphor is set as Sorgq enhancement coefficient β and the reproduced image signal S', the calculation expressed by the calculation formula %) is performed to obtain the above ultra-low space. In a radiation image processing method that emphasizes frequency components higher than the frequency,
The unsharp mask signal Sus is converted into an original image signal S org at each scanning point within a rectangular unsharp mask surrounded by two sides parallel to the main scanning direction and two sides parallel to the sub-scanning direction. This is a radiation image processing method characterized by simple addition and averaging.

The apparatus of the present invention also includes an excitation light source for scanning a stimulable phosphor to stimulate the radiation image stored therein, and a photodetector for detecting the emitted light and converting it into an electrical signal. The output of this photodetector is simply added and averaged over the range of a rectangular non-sharp mask surrounded by two sides parallel to the main scanning direction and two sides parallel to the sub-scanning direction. From a circuit for obtaining an unsharp mask signal Sus corresponding to an extremely low spatial frequency, this unsharp mask signal 5uss, an original image signal Sorgs which is the output of the photodetector, and an emphasis coefficient β, when the reproduced image signal is S' , a radiation image processing apparatus includes an arithmetic device that performs an arithmetic operation expressed by the arithmetic expression S'-Sorg+β(Sorg-5us).

In the present invention, the unsharp mask signal Sus corresponding to ultra-low frequencies refers to an unsharp image (
This refers to the signal at each scanning point of the "unsharp mask" (hereinafter referred to as the "unsharp mask"). As this unsharp mask, the modulation transfer function is 0.
0.5 or more when the spatial frequency is Ol cycle/#,
and 0.5 cycles/mrs spatial frequency.
5 or less, or where the integral value of the modulation transfer function with 0.01 as the lower end in the spatial frequency range of 0.01 to 0.5 cycles/# is 0.01 to 10 cycles/lN11. A value that is 90% or more of the integral value of the modulation transfer function is used.

In addition, as disclosed in the patent application (4) dated November 22, 1978 (applicant: Fuji Photo Film Co., Ltd.), when the modulation transfer function is 0.02 cycles/hazy spatial frequency, .5 or more and 0.15 cycles/m
Using an unsharp mask that is less than 0.5 at a spatial frequency of m significantly improves the diagnostic performance.

Here, the above-mentioned unsharp mask has a modulation transfer function value of 0.
If we define the spatial frequency of 5 as fc, it is equivalent to saying that the modulation transfer function f is preferably in the range of 0.01 to 0.5 cycles/m, preferably 0.02 to 0.15 cycles/m. It is.

In the present invention, the original signal includes a signal that has been processed by means commonly used in the optical industry, that is, a signal that has been subjected to nonlinear amplification such as logarithmic amplification for band compression and nonlinear correction. Needless to say.

In the present invention, when simple averaging is performed using a rectangular unsharp mask, in other words, when the unsharp mask is rectangular and the weight of the original signal (Sorg) of the pixel included in the mask is constant, fゎ is 0. The regulation of O1 to 0.5 cycles/# (preferably 0.02 to 0.15 cycles/rIIrR) theoretically reduces the length of one side of the rectangular non-sharp mask to 6h+m-1.2mIII (preferably, 30#III to 4Inm). Note that even when the shape of the non-sharp mask is rectangular, the length of each side only needs to be within the above range, and for example, a rectangular mask with a large aspect ratio is effective for image processing of linear tomography.

In the present invention, the enhancement coefficient β includes cases where it is a constant and cases where it is a function of the original image signal (S org) or the unsharp mask signal (S us), but especially in the latter case, that is, the enhancement coefficient β It is preferable to change the value according to the original image signal (Sorg) or the unsharp mask signal (S us ), since the diagnostic performance can be further improved.

Furthermore, depending on how the emphasis coefficient β and the unsharp mask signal (S us ) are selected, the maximum value (B) of the modulation transfer function of the system that provides the visible image created based on the signal emphasized by the present invention can be determined. Value of modulation transfer function near zero frequency (A)
The ratio (B/A) changes, but when B/A<1.5, there is almost no difference in diagnostic performance compared to conventional X-ray photography. Furthermore, when performing the processing of the present invention with the emphasis coefficient β as a constant, if B/A exceeds 6, unnatural image parts may appear due to excessive emphasis, or the image may appear white or black. This is not desirable as it often causes a partial appearance and interferes with the diagnosis. On the other hand, when the emphasis coefficient β is changed according to the original image signal S org or the unsharp mask signal Sus, the preferable range of B/A (in this case, B/A also changes according to Sorg or Sus However, B/A is its maximum value.) is expanded,
Even if B/A exceeded 6, when this was 10 or less, the above-mentioned false image was not noticeable. Furthermore, B
When the value of /A was set in the range of 2 to 5,5 when β was fixed, and in the range of 2 to 8 when β was variable, the diagnostic performance was significantly improved.

Further, the emphasis coefficient β is set so that B/A is within the above range, but B/A changes slightly depending on the shape of the sharpening mask β, that is, Sus, but B/A− 1,
5 to 10, when using a simple average mask,
This is synonymous with setting β to 0.4 to 8.

In the present invention, in addition to the above operations, smoothing processing can also be performed. In general, in the high frequency region, there is a lot of noise and it is often difficult to see, so it is often preferable to perform further smoothing processing to further improve diagnostic performance. Preferably, the smoothing process is such that the modulation transfer function is 0.5 or more when the spatial frequency is 0.5 cycles/# and is 0.5 or less when the spatial frequency is 5 cycles/#. What kind of smoothing processing is preferable? For example, when interpreting a relatively low frequency shadow such as a chest tomographic image, it is preferable to remove as much noise as possible.
On the other hand, when it is necessary to trace fine blood vessel shadows containing high frequency components, such as in angiography images, too strong a smoothing process may obscure the desired shadows, which is undesirable. Although it varies depending on the site, symptoms, purpose of examination, etc., according to the research of the present inventor, performing the above-mentioned smoothing process has the effect of improving diagnostic performance for almost all X-ray images. There was found. Furthermore, this smoothing process is equally effective whether it is applied to S′ after the ultra-low spatial frequency processing of the present invention or to the original image signal S org. It is acknowledged that there is.

Further, in the present invention, gradation processing may be performed in addition to frequency emphasis processing using a non-sharp mask. Ultra-low frequency processing has a relatively small effect on diseases where luminance changes slowly over a large area, such as lung cancer and breast cancer.
No. 8740, Special Publication No. 62-53179, No. 63-28
It is desirable to use the gradation processing disclosed in No. 585 or the like in combination. In this case, gradation processing is performed before and after ultra-low frequency processing,
This can be done either way.

In the present invention, a phosphor refers to a phosphor that is irradiated with the first light or high-energy radiation and then is stimulated (excited) thermally, mechanically, chemically, electrically, etc. It refers to a so-called photostimulable phosphor that re-emits light corresponding to the amount of radiation irradiated, and those having a photostimulable emission wavelength of 300 to 500 nIIl are particularly preferable, such as rare earth element-activated alkaline earth metals. Fluorohalide phosphor [specifically, Special Publication No. 60-42
It is described in the specification of No. 837 (Bal-1-F
1Mg −, Ca y ) FX: aE′u2+ (however, X
is at least one of C9J and B'', and
and y is 0, x+y≦0.6 and xy≠0, and a
is lO°6≦a≦5×l0-2)
4333 (Bat-x, M
n, ) F X : y A (However, Mn is Mg, C
a, at least one of Cr, Zn and Cd, X
is at least one of C9J, Br and ■, A is Eu, Tb, Ce, Tra, Dy, Pr, Ho, Nd,
At least one of Yb and E", X is 0≦X≦
o, e, y are 0≦y≦0.2), etc.; Zn S:C described in Japanese Patent Publication No. 60-9542
u, PbSBaO・x AQ, 203: Eu (
However, 0.8≦X≦10) and M O・x Sl 02 :A (However, Mn is M
g.

Ca, Sr, Zn, Cd or Ba, A is Ce,
Tb, Eu, Tm, Pb, Tfl, Bi or Mrl, and X is 0.5≦X≦2.5); and LnOX:
A (However, Ln is at least one of La, Y, Gd and Lu, X is at least one of C9J and Br, A is at least one of Ce and Tb,
X is 0<x<0.1); and the like. Among these, rare earth element-activated alkaline earth metal fluorohalide phosphors are preferred, and among these, barium fluorohalides shown as specific examples are particularly preferred because of their excellent stimulable luminescence.

Furthermore, if the phosphor layer of a stimulable phosphor plate made using this stimulable phosphor is colored with a pigment or dye, the sharpness of the final image will be improved and favorable results will be obtained. (In the present invention published in Japanese Patent Publication No. 59-23400,
Laser light with good directivity is used as excitation light for reading out the radiation image accumulated on the stimulable phosphor plate. As the excitation light source of the laser light, in order to easily separate it from the emitted light and increase the S/N ratio, a laser beam with a wavelength of 500 to 800 nm s is used.
Preferably, a laser that emits light in the range of 600 to 700 nm, such as a He-Ne laser (B33n11) or a Kr laser (647nI11), is preferable;
Excitation light sources other than those described above can also be used by using a filter that cuts light other than m.

A radiographic image subjected to image processing according to the present invention is input to a CRT and reproduced as a visible image on the CRT, thereby enabling CRT diagnosis. In addition, after displaying the above-mentioned radiographic image on a CRT etc. and observing it, the radiographic image is displayed on a silver halide photographic film, a diazo film,
It may also be optically recorded on a recording material such as an electrophotographic material.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in detail below based on an X-ray image information recording and reproducing system that is an embodiment thereof.

FIG. 1 shows the process of creating a reproduced image.

When X-rays are emitted and irradiated onto a human body, the X-rays that pass through the human body enter the phosphor plate. This phosphor plate stores the energy of the x-ray image at the phosphor trap level.

After taking the X-ray image, the phosphor plate is scanned with excitation light having a wavelength of 500 to 800 nm to excite the accumulated energy from the traps, causing them to emit light in the wavelength range of 300 to 500 nm. This emitted light is measured with a photodetector, such as a photomultiplier tube or a photodiode, which receives only light in this wavelength range.

After reading the X-ray image, the output signal of the photodetector is nonlinearly amplified and then converted into a digital signal by an A/D converter and stored on a magnetic tape.

The digital signals of each section stored on this magnetic tape are read out to a calculation device, such as a minicomputer, and after determining SO8, the above-mentioned calculation of S'-8org+β(Sorg-5us) is performed.

The Sus is such that the modulation transfer function is 0.5 or more at a spatial frequency of 0.01 cycles/mm and 0.5 or less at a spatial frequency of 0°5 cycles/m. Must be specified. Furthermore, when calculating the above equation, it is necessary to specify the emphasis coefficient β. These values may be specified individually from the outside, or several types may be determined for each part of the human body or for each case, and these values may be stored in the memory of the computing device.

Smoothing processing is performed on the S' to reduce frequency components higher than ultra-low spatial frequencies. This smoothing process can reduce noise without damaging information necessary for diagnosis.

As described above, the present invention is characterized in that the unsharp mask signal Sus is obtained by simple averaging of the signals within the mask using a rectangular unsharp mask, and this method is extremely simple. The unsharp mask signal Sus can be obtained by a similar method. This is a common advantage when signal processing is performed in either digital or analog format.In practice, this method makes it possible to obtain the unsharp mask signal Sus in an extremely short time. This is the first time that unsharp mask processing using this method can be implemented in a practical sense.

That is, for example, the signal S org(i , j
) is multiplied by the weighting coefficient alj, the unsharp mask signal 5us (l is 5us (I bar Σaij @ S org (ij) I mock j (1, j are pixels indicating the coordinates of each scanning point) The numbers l and J are the numbers indicating the coordinates of the unsharp mask signal, and are determined by the calculation Σaij-1). Therefore, the number of calculations is approximately N2 times of multiplication and the same <N2 times of addition. When the number of pixels in the non-sharp mask increases, N (number of pixels representing the length of one side of the non-sharp mask) increases, it takes a considerable amount of time to obtain the non-sharp mask signal Sus. The size of the sharpening mask is 6 mm x 6 mm, and the size of the pixel (0.1 mm x 0
．． l mm>, it takes 3600 multiplications and 3
The addition must be repeated 600 times, for example g b
When calculating only with software using a tt microcomputer, for example, multiplication is 3□. , addition is 5#.

, C, it takes 3□llc X a800 + 511ute
3800: 11am, which is completely impractical.

In contrast, according to the present invention, simple arithmetic averaging is sufficient, so multiplication is not necessary, and calculation time can be greatly reduced.

For example, in the above example, each point is 1811a*c. Furthermore, it is possible to simplify various calculations as described below, and depending on the calculation algorithm, the number of calculations can be drastically reduced to only 4, and it is also possible to obtain the unsharp mask signal Sus within several tens of Masc. Therefore, the practical effects of the present invention are remarkable. That is, since 5us(IJ) is obtained by 5us(IJ)=(ΣS lj), Sus can be obtained by only N2 additions and one multiplication.

In more detail, the size of the unsharp mask is N1 in the main scanning direction and N2 in the sub-scanning direction, and the unsharp mask signal 5u is
s(IJ) is expressed as 5us(IJ)=(ΣS 1j) N1 xl'J2, and can be calculated simply by performing NI XNz additions and one division.

Furthermore, by devising the calculation procedure using various algorithms described below, it is possible to reduce the average number of calculations to obtain one non-sharp mask signal to just four.

Hereinafter, an example of an algorithm (digital method) that particularly simplifies calculation for obtaining the above-mentioned unsharp mask signal Sus will be explained.

As shown in FIG. 2, consider a rectangular non-sharp mask M (indicated by a thick solid line) surrounded by two sides parallel to the main scanning direction and two sides parallel to the sub-scanning direction.

For simplicity, this mask M is a square, and the length of its negative side is expressed as the number of pixels. (N is a positive odd number) Second
In the figure, the signal value of the scanning point (pixel) that S' IJ attempts to obtain by image processing (the above-mentioned calculation formula S'-3org
+β(Sorg-Sus) final signal value)
, the signal value of pixel PIJ input at the time of interest when SIJ is at the tip of the mask in the scanning direction, T1. is the sum of the signal values of N2 pixels in the mask M, that is.

Here, the signal value SLI of the pixel PIJ of interest is first stored at the corresponding address of the pixel signal S. Each address requires the number of bits (for example, 8 bits) that can represent the signal value of the pixel.

Next, the sum CIJ of signals for N pixels in the main scanning direction is calculated by
The sum CI-1,J of the signal values of the pixels of interest, the signal value 5t-N of the pixel N places before the pixel of interest PIJ, and the pixel of interest P1. Signal value S of J 1. According to J, the calculation formula C1, J = Cl-1, J +S1. J””5
l-N.

, you can find it from . Then, this sum C1J is stored at the corresponding address of the sum C of signals of the columns of pixels in the main scanning direction. Each address requires a sufficient number of bits to perform this operation without overflowing, and this number of bits depends on N.

Next, the total sum T, , of the signal values of N2 pixels in the mask Ml, J is determined. This is the pixel P1.
The total sum T I, J-1 of the signal values of the pixels in the mask M1, J-1 located one column back in the sub-scanning direction from the mask M including J at the tip, and the mask M1. The sum C of the signals of the last column of J-1 (that is, the column that is no longer included in Ml, )
1, J-N and the pixel of interest P1. ．． By the sum C1,J of the signals of the column including at the tip, the calculation formula T1. J-T
1. It can be determined from J-1+C1, J C1, and J-N. Then, this value T1. J is stored at the corresponding address of the sum T of the signals of the pixels in the non-sharp mask. This T
1. Since J corresponds to N2 times the unsharp mask signal Sus, after obtaining this TIJ, the aforementioned unsharp mask process can be performed using the arithmetic expression.

The memory capacity required for the above calculation will be explained next.

FIG. 3(a) shows a memory for SIJ, which requires memories for all the number of pixels required in the main scanning direction, and N''. memories in the sub-scanning direction.

For example, one memory only needs to have a capacity of 8 bits.

FIG. 3(b) shows a memory for CIJ, and it is sufficient to have the same number of memories in the main scanning direction as the memories for SIJ, and N+1 memories in the sub-scanning direction.

This memory requires 2-3 times the number of bits as the memory above. FIG. 3(C) shows a memory for TIJ, which requires the same number of memories as the above two memories in the main scanning direction, but only needs to have two memories in the sub-scanning direction.

FIG. 4 shows an example of a circuit block that performs the above calculation, in which a signal is sent from the gate 11 to which the pixel input signal SIN is input to the memory 12 having the above capacity, and the memory 12
The calculation circuit 13 performs calculations based on the stored values. These gates11. The memory 12 and the arithmetic circuit 13 are operated by a control circuit 14.

The calculation result by the calculation circuit 13 is output from the gate 11 via the memory 12 as a pixel output signal 80υT.

According to the above calculation method, the calculation for obtaining the unsharp mask signal Sus is extremely simplified, and the apparatus therefor is also extremely simplified. This is based on the fact that, according to the method of the present invention, an unsharp mask signal is obtained by simply averaging signal values of pixels within a rectangular mask.

That is, according to the simple averaging method of the present invention, an extremely simplified algorithm such as the one described above becomes possible, and calculations can be performed extremely easily.
The radiation image processing that is the object of the present invention can be realized very easily.

In addition, in the arithmetic circuit that implements the above algorithm,
As shown in FIG. 5, the three types of memories 18, 19, and 20 can be arranged as a series of memories with consecutive addresses, but the three types of memories 15, 16, and 17 can also be connected to an address bus as shown in FIG. By dividing the data bus and making it possible to access three memories simultaneously, calculation time can be further reduced.

The control circuit and the arithmetic circuit are each made of dedicated hardware such as PLA (Progra+nable).
LogIc array), random logic (
randoIIlogle) etc. may be used. Also,
A microcomputer, minicomputer, etc. may be used for these circuits, or a high-speed microcomputer (for example, bit slice type) may be used for the control circuit.
A dedicated circuit may be used as the arithmetic circuit.

This can be determined by selecting an appropriate one depending on the required calculation speed.

An example of an algorithm that can reduce memory capacity even further than the above algorithm is shown in Figure 7 below.
This will be explained with reference to FIGS. 8 and 9.

In this algorithm, after storing the pixel of interest, that is, the signal value S1 of the non-sharp mask M+ and the leading pixel Pi in the corresponding address of the memory for S,
The sum EIJ of the signals of N pixels of IJ, that is, EIJ-ΣS.

1-J-N+1 is calculated and the value is stored in the corresponding address of the memory for E. This is done by the arithmetic expression %expression %.

Using these stored values, the unsharp mask signal S
TIJ corresponding to N2 times us is calculated.

This TIJ is determined by the arithmetic expression TIJ-"T, -, , J+E, , -E, -N, J.

However, with this method, calculation is not possible when the main scanning returns from the right end to the left end, so at this time, the sum R of the first N signals SIJ in the main scanning direction, that is, R4 - Σ S1j, is calculated and this is stored in the memory for R. Store in the corresponding address. This R1 is, for example, N-5 as shown in FIG.
When , R5 is the sum of Sl, 1 to S5.1, and R5 is Sl, , to S91. is the sum of S.

5 to S6. , R5 does not change even if the output changes.

Therefore, when main scanning returns from the right end to the left end, the above R,
T1. is required.

Using the TIJ obtained in this way, the above-mentioned unsharp masking process can be performed using an arithmetic expression.

In this algorithm, the memory for the signal value Slj of each pixel is as shown in FIG. However, R, E,
Figure 8 (b), (c), etc. are used as memories for T.
As shown in (d), R and E have N+1 memories in the main scanning direction and 1 memory in the sub-scanning direction, and T has a small capacity of 2 memories in the main scanning direction and 1 memory in the sub-scanning direction. memory can be used. For example, 8 bits can be used for each address in the S memory, but R, E.

For example, a 16-bit memory (depending on the size of N) is required for T. In this algorithm, the memory for S, which requires a small number of bits, is made large, and the other memories, which have a large number of bits, are made small instead, so the overall memory capacity can be significantly reduced. Therefore, the capacity of the memory shown in FIG. 8 can be made much smaller than that shown in FIG. 3, which has a great effect on the simplification of the device.

In addition, the positive odd number N in the above two methods is preferably about 10 pixels/mm in order to obtain the image accuracy necessary for diagnosis, and in that case, it is 601 to 11, preferably 301.
A size in the range of ~39 is preferable.

Both of the above two algorithms utilize a method of digitally processing signals, but they integrate the signal at each scanning point in the main scanning direction in an analog manner, and store the integrated value in memory. , even if this is numerically integrated, Su
s can be obtained. In this case, instead of using a digital circuit, analog values are integrated and added for each pixel, so N
However, the following method can advantageously reduce the number of integrators to one.

That is, the analog output Sorg of each scanning point is divided into two,
One is a delay circuit (the delay time (T) is the scanning time of one pixel (
τ) x number of pixels in the main scanning direction of the non-sharp mask (N), that is, T - τ
) can also be used to obtain the value obtained by integrating. This value corresponds to C1 and J in Figs. 2 and 3, and by digitally adding these values in the sub-scanning direction, Tol is obtained, from which the unsharp mask signal Sus can be obtained. . This is also a fast and easy calculation method.
It is suitable as an analog method.

Note that the unsharp mask signal 5us (I) is calculated only for the signal S1j at the scanning points within the mask range of N1-1 N1-1 with one scanning point (i, j) as the center. Therefore, an unsharp mask signal centered on a scan point at the edge of the image cannot be determined because there is no signal outside the edge.

As a method for processing this edge, assume that the SIJ value at the outermost circumference extends outward infinitely, store the value at the outermost circumference in memory, and use this value as a signal outside the edge. It is natural and advantageous to do so. Alternatively, the area outside the outermost periphery may be processed as black or white, or may be processed as a constant intermediate value between black and white.

In the above method, only one unsharp mask is used to perform unsharp mask processing, but it is also possible to use two unsharp masks of different sizes to provide stages in frequency emphasis. In this case, we will perform the calculation expressed by the formula % formula %), but if we rewrite this formula, S' = Sorg + (β + α) (Sorg - (βSus 1 + asus Z ) 1
It can also be expressed as β+α, which corresponds to an operation similar to the above-mentioned operation. Non-sharp mask 5
When S us2 is smaller than usl and the emphasis coefficient α is positive, the graph of the modulation transfer function has an additional peak at the higher part of the emphasized frequencies, and when α is negative, the emphasis coefficient α is positive. The higher part becomes lower in stages. The former is particularly suitable for bone regions, angiography, double gastroscopy, etc., and the rear part is particularly suitable for chest tomography, cholecystography, hepatography, plain abdominal imaging of two heads, etc.

In addition, in the above explanation, the original image signal Sorg is
It also includes a signal after band compression such as logarithmic transformation and nonlinear correction. Practically speaking, since the output of the photodetector is subjected to signal processing, it is desirable to perform band compression such as logarithmic transformation. It goes without saying that, in principle, it is also possible to use the output of the photodetector as it is as Sorg for subsequent processing. Also, theoretically, this mask calculation should calculate the average energy, but
According to the inventor's experiments, when obtaining this unsharp mask signal, even if the average value was calculated using a value corresponding to the logarithmically compressed density, the result did not change. This is practically advantageous in terms of processing.

Hereinafter, the arithmetic processing using the above-mentioned non-sharp mask will be explained in more detail with reference to FIG. 1O.

FIG. 10(a) shows the cycle number response when the accumulated image on the phosphor is sampled by IO pixel/unit. It is known that this curve becomes a 5ine curve when a rectangular aperture is used as the aperture of the photodetector, and becomes a Gaussian distribution curve when a Gaussian distribution aperture is used.

In Fig. 1O(b), the modulation transfer function is 0. Ol cycle/m
This shows a rectangular non-sharp mask using a mask that is 0.5 or more when the spatial frequency is 0.5 cycles/mm and 0.5 or less when the spatial frequency is 0.5 cycles/mm.

When the image on the phosphor is sampled at IO pixel/#, an unsharp mask is created by taking a simple average of approximately 63 pixels x 63 pixels (this is called "unsharp mask size N - 634"). This is equivalent to scanning the image on the phosphor with a large-sized light beam of 6.3 #1Ill x 6.3m.

FIG. 10(e) is a graph showing the modulation transfer function after the calculation of (Sorg-9us).

The solid line (1) in FIG. 10(d) shows the calculation result S'. Here, β is fixed at "3".

As a result of the above calculation, the maximum value (B) of the modulation transfer function of the enhanced image signal is approximately 4.6 times the modulation transfer function (A) near zero frequency.

The dotted line (n) in FIG. 1θ(d) shows the modulation transfer function when S′ in FIG. 1O(d) is subjected to smoothing processing of 5 pixels×5 pixels.

Figures 11A to 11D show the emphasis coefficient β as the original image signal (Sorg) or the unsharp mask signal (Sus).
This shows an example in which the values are continuously changed according to the .

Fig. 11A shows a flat type with β constant, Fig. 11B shows a monotonically increasing type (β'≧0), Figs. 11C and 11D both include the case where β' becomes 0, and Fig. 11C shows the low brightness emphasis type, and the first LD shows the medium brightness emphasis type, and these include a stepped change (curve a) and a curved change (curve b).
).

As shown in FIG. 11B, by changing β monotonically, it is possible to prevent false images that tend to occur due to frequency emphasis. As an example, when performing the frequency processing on an X-ray image of the stomach using a barium contrast agent while fixing the enhancement coefficient β, the boundary of a wide low-intensity region containing a large amount of contrast agent is emphasized more than necessary, resulting in a double-contoured false image. Instead, the enhancement coefficient β is made variable, that is, β is made smaller in the low-brightness region where a large amount of contrast agent is present.
By increasing β in a high brightness area such as a gastric subdivision, the occurrence of the double contour can be prevented. As another example, in the case of frontal chest photography, if β is fixed, noise will increase in the low-brightness region of the spine and heart, and in extreme cases, details will be washed out (this is visually very noticeable, adversely affect diagnostic performance). Similarly, if β is made small in the low brightness areas of the spine and heart, and β is made large in the high brightness areas of the lung fields, the increase in noise and white areas described above can be prevented.

The low brightness enhancement shown in FIG. 11c is suitable for cases where diagnosis of low brightness areas is particularly important and the low brightness areas do not occupy a large portion of the entire image. For example, this applies to angiography and lymphangiography, and in these radiographic images, it is desirable to greatly improve the sharpness of the desired area even if the noise increases slightly, so this low brightness enhancement greatly improves diagnostic performance. improve.

Furthermore, the medium brightness enhancement in Figure 11D is used when low brightness areas and high brightness areas occupy a considerable portion of the entire image, the area in parentheses is not important for diagnosis, and the medium brightness areas are particularly important for diagnosis. Are suitable. For example, cholecystography and hepatography correspond to this case, and if noise or gas areas are emphasized in these radiographic images, it will interfere with diagnosis. It is desirable to emphasize.

In any of the above examples, if the emphasis coefficient β is fixed to a small value and frequency processing is performed, various false images will not occur, but the gastric subarea, which makes an important contribution to diagnostic performance, will not be generated. The contrast of blood vessels in the lung field and contrast-enhanced vessels does not improve, and diagnostic performance does not improve. By continuously changing the enhancement coefficient β in accordance with the brightness of the image on the phosphor in this way, an image with improved diagnostic performance can be obtained while preventing the generation of false images.

FIG. 12 shows an example of how to increase β, and the minimum brightness S is determined from the histogram of the image on the phosphor. and the maximum brightness S1 are determined, and β is changed approximately linearly between these values. So and S are determined depending on the type of X-ray image to be processed, and for example, the minimum and maximum brightness may be the brightness values when the integral histogram is 0 to 10% and 90 to 100%, respectively.

FIG. 13 and FIG. 14 show examples of how to change β in low brightness emphasis and medium brightness emphasis, respectively.

In FIG. 13, β is the maximum value βma between brightness A and B.
x to the minimum value βmin. In other words, the emphasis coefficient is increased (βmax)L in the low luminance region (from S akin to A), and the emphasis coefficient is increased (βmax) L in the low luminance region (from B to S
(up to Lllax), it is small (βm1n). Brightness A is the minimum brightness (SmIn) and the maximum brightness (
The difference (ΔS
) plus 0.2 to 0.5 times [Sm1n +
(0,2-0.5)ΔS] is good, and the brightness B is the same <0.
The size obtained by adding 7 to 1 times [Dn+1n + (0,7 to
1) ΔD] is good.

In FIG. 14, β increases from the first minimum value (βm1ni) to the maximum value (βll1ax) between brightness A and B, and increases from the maximum value (βWaX) to the second minimum value between C and D. (βm+1n2). That is, the low brightness area (from S mIn to A) and the high brightness area (from D to S
1iax), the emphasis coefficient is reduced (βm
1n1. βl1in 2) L, , medium brightness region (
from B to C) is large (βff1ax)
are doing.

Here, the first minimum value (βl1ini) and the second minimum value (
βm+1n2) may be equal. - In the case of the dashed dotted line, β increases between brightness A and E, and decreases between brightness E and D. Brightness A is the minimum brightness (Smin) plus O~0.2 times the difference (ΔS) between the maximum brightness (SIllax) and the minimum brightness (SIIlln) [Sm1n
+ (0 to 0.2) ΔS], and the brightness B is calculated from the average brightness S 1n + S ll1ax or statistical average value) to the difference (ΔS) from 0 to 0.
The size obtained by subtracting 2 times [5 - (0 to 0.2) ΔS, luminance E is the average luminance (S), and luminance C is the difference (ΔS) from the average luminance.
) plus 0 to 0.2 times [S+(0 to 0.2)
ΔS, brightness is calculated from the maximum brightness (S + HX) by the difference (
ΔS) minus 0 to 0.2 times [Smax −(
0 to 0.2) ΔS] are respectively desirable.

In addition, in the calculations in FIGS. 13 and 14 above, the maximum brightness (Sa+ax) and the minimum brightness (Smin) both correspond to the maximum and minimum in the target substantial image, and It is possible that a portion has a brightness greater or less than this. Note that depending on the case, the maximum and minimum values among all the screens may be simply taken.

In addition, in experiments conducted by the present inventors, the effect of changing β with the original image signal of the image on the phosphor and the case with changing β with a non-sharp mask signal was different from the difference between lines, etc. Ta.

In addition to the frequency emphasis processing using a non-sharp mask as described above, gradation processing can also be used in combination.

When gradation processing is performed before ultra-low frequency processing, A/D conversion is performed after gradation processing is performed using a nonlinear analog circuit. If it is performed after A/D conversion, digital processing can also be performed using a minicomputer. After ultra-low frequency processing, digital processing or D/A conversion converter analog processing is performed. The data, which has been subjected to frequency enhancement and gradation processing as necessary, is recorded on a magnetic tape. The data on this magnetic tape is sequentially read out, converted to an analog signal by a D/A converter, and amplified by an amplifier.
It is input to a CRT and reproduced as a visible image on the CRT,
CRT diagnosis becomes possible. Further, this data is further input to the recording light source as required. Light generated from this recording light source passes through a lens and is irradiated onto a recording material, such as a photographic film, so that a radiation image is reproduced on the photographic film. Therefore, in that case, diagnosis can be made by observing the image on the film. Note that when reproducing and recording images on a CRT or photographic film, a reduced image can be obtained by recording at a higher sampling frequency than during input scanning. For example, if the input system scans at 10 pixels/mn and the output system scans at 20 pixels/mn, the image will be reduced to 1/2. In an image reduced to 172 to I/3 in this way, the frequency components considered necessary for diagnosis are close to the frequency range with the highest visibility, so the contrast appears to be visually higher, making it very easy to see.

The present invention is not limited to the embodiments described above, and various configuration changes are possible.

The image on the phosphor can be read out by a method of setting the phosphor on a rotating drum, a method of two-dimensional scanning in a plane, or an electronic scanning method such as a flying spot scanner. Further, the unsharp mask calculation can be performed by digitally processing only the sub-scanning direction by unsharpening the analog signal using a low-pass filter in the main scanning direction before A/D conversion. Further, the above calculation may be performed offline after all data is stored on the magnetic tape, or may be partially stored in a core memory and sequentially processed online.

Example For a total of 200 cases for the sites shown in Table 1,
Images recorded directly on conventional X-ray photographic film and images read out from phosphors and created using ultra-low frequency processing according to the present invention were compared to examine improvements in diagnostic performance for major parts of the human body.

The reproduced image in Table 1 is calculated using the spatial frequency at which the modulation transfer function in the side direction of the rectangle is 0.5 by fixing the emphasis coefficient β to 3 and using a simple average of image signals in a rectangular area as a non-sharp mask. It was created by changing fc in six ways.

Since it is virtually impossible to confirm the existence and degree of improvement in diagnostic performance using physical evaluation values of ordinary photography (e.g., sharpness, contrast, graininess, etc.), The evaluation was based on subjective evaluations by a total of 20 radiation reading experts: 12 clinicians, and 4 radiology technicians.

The evaluation criteria were as follows.

+2 = Lesions that could not be seen with the conventional X-ray film method became visible, lesions that were extremely difficult to diagnose became easier to see, and diagnostic performance was clearly improved.

+1 = Lesions that are difficult to diagnose using conventional X-ray film methods are now easier to see, improving diagnostic performance.

0: Compared to the conventional X-ray photographic film method, it is easier to see, but no particular improvement in diagnostic performance is observed.

-1: Although there were areas where diagnostic performance improved, there were also areas where it was difficult to diagnose.

-2: There were no areas where diagnostic performance improved, and some areas were difficult to diagnose.

FIG. 15 shows the results of averaging the expert evaluation values for a total of 200 cases at the sites listed in Table 1. The curve shown in FIG. 15 is obtained by further averaging these averaged evaluation results.

From Figure 15, the spatial frequency f at which the diagnostic performance is particularly improved
The range of c was found to be between 0.02 and 0.15 cycles/#.

In addition, through this experiment, we determined the value of fc that gives the best evaluation and the evaluation value at that time, in other words, the peak position, depending on the evaluator's preference, the area to be imaged, the case, and the purpose of imaging (screening or detailed examination, etc.) Although it varies considerably depending on the presence or absence of other clinical test findings, etc., it has been found that the range of fc in which the effects of processing according to the present invention are recognized has relatively little variation for all images.

In addition, for a total of 20 columns of representative cases shown in Table 1, fc
Images were created in the same manner by varying the constant B/A fixed at 0.05 cycles/ml1, and similar evaluations were performed by a total of 20 radiology interpretation experts.

The 16th edition shows the average value of the evaluation for each case.
It is a diagram. Curve a in Figure 16 is the maximum B/A when β is kept constant regardless of the signal of the original image, and curve a is the maximum B/A when β is continuously changed depending on the signal of the original image.
This is the result for the value of . If β of curve a is constant, B/A
When the value becomes 6 to 7 or more, false images become noticeable and the evaluation becomes less than 0, but if β is made variable, false images are removed and 1.5≦B
The evaluation was 0 or more in the range of /A≦10. Almost similar improvements in diagnostic performance were observed for other types of cases.

In addition, the improvement of diagnostic performance by combining ultra-low frequency emphasis with other processing (change of emphasis coefficient β, gradation processing, reduction, smoothing processing) was carried out for the various cases mentioned above,
In both cases, the results showed that the diagnostic performance was further improved.

Since the present invention having the above configuration emphasizes the frequency response from the very low frequency region, the frequency region important for diagnosis is greatly emphasized. Therefore, contrast is improved and diagnostic performance is improved. Furthermore, by changing the degree of enhancement depending on the brightness signal, shape, etc., it is possible to prevent the generation of false images and to prevent diseases important for diagnosis from becoming difficult to see.

Furthermore, since high-frequency components are not emphasized, there are fewer noise components (and the image becomes smoother).

As a result, an image that is easy to view can be obtained.

All of these image processing effects are achieved by considering how to approach the optimal frequency of the modulation transfer function for human vision. Reduction is particularly effective.

[Brief explanation of the drawing]

FIG. 1 is a flow sheet showing the process of processing an X-ray image according to the present invention. FIG. 2 is a diagram showing a non-sharp mask, pixels, etc. on an image to explain one algorithm for calculating a non-sharp mask signal. FIG. 3 is a diagram showing the memory capacity when the above algorithm is used. FIG. 4 is a block diagram showing an example of the configuration of a circuit that performs calculations using the above algorithm. FIGS. 5 and 6 are diagrams showing examples of changes in the memory configuration in the above configuration. 7 and 9 are diagrams showing a non-sharp mask, pixels, etc. on an image to explain another algorithm for calculating a non-sharp mask signal. FIG. 8 is a diagram showing the memory capacity when this algorithm is used. FIG. 1O is a graph showing the steps of frequency enhancement. FIG. 11 is a diagram showing an example of changing the emphasis coefficient β according to the luminance. FIG. 12 is a graph showing an example of a combination of emphasis coefficient β and original image signal Sorg. FIGS. 13 and 14 are diagrams showing an example of a specific method for changing the emphasis coefficient β depending on the image signal. FIG. 15 is a graph showing the results of diagnostic performance evaluation in the example. Figure 16 shows the maximum modulation transfer function (B) and the modulation transfer function (A) near the zero spatial frequency in the emphasized copy photograph.
) is a graph showing the relationship between the ratio B/A and the evaluation of diagnostic performance. Figure 1 Figure 2 Figure 3 Figure 1! I! :” Ritsuka IL Power nI:tflR+t*e!
+1゛ヱ Scanning direction l ≠ 1 stroke 1 dazzle figure figure figure figure figure figure /112 number (cycle 1 /rnrn) figure org figure figure +14 figure figure 1 old figure figure 1c figure D Figure βTomiC0n5t β“2゜β”S. l! O A request for examination of the application is submitted at the same time as the withdrawal.

Claims (1)

[Claims] 1) A stimulable phosphor is scanned with excitation light to read radiation image information recorded on the phosphor material, convert it into an electrical signal, and then reproduce it as a visible image. , the unsharp mask signal Sus corresponding to the very low spatial frequency at each scanning point
, and when the original image signal read from the phosphor is Sorg, the emphasis coefficient is β, and the reproduced image signal is S′, perform the calculation expressed by the formula S′=Sorg+β(Sorg−Sus). In the radiation image processing method for emphasizing frequency components higher than ultra-low spatial frequencies, the unsharp mask signal Sus is formed by two sides parallel to the main scanning direction and two sides parallel to the sub-scanning direction of the scanning. A radiation image processing method characterized in that original image signals Sorg at each scanning point within an enclosed rectangular area are obtained by simple averaging. 2) The radiation image processing method according to claim 1, wherein the visible image is reproduced on a CRT. 3) The rectangular unsharp mask for obtaining the unsharp mask signal Sus has a rectangular side length of 60 mm to 1.
The radiation image processing method according to claim 1 or 2, characterized in that the size is in the range of 2 mm. 4) The radiation image processing method according to claim 3, wherein the rectangular non-sharp mask has a size such that the length of one side of the rectangle is in the range of 30 mm to 4 mm. 5) The radiation image processing method according to claim 1 or 2, wherein the emphasis coefficient β is a constant. 6) The fifth aspect of the present invention is characterized in that the maximum modulation transfer function of the system that provides the image enhanced by the arithmetic expression is 1.5 to 6 times the modulation transfer function near the zero spatial frequency. The radiation image processing method described in Section 1. 7) The radiation image processing method according to claim 1 or 2, characterized in that the emphasis coefficient β is changed according to the value of the original image signal or the non-sharp mask signal. 8) Claim 7, characterized in that the maximum modulation transfer function of the system that provides the image enhanced by the arithmetic expression is 1.5 to 10 times the modulation transfer function near the zero spatial frequency. The radiation image processing method described in Section 1. 9) In addition to emphasizing ultra-low spatial frequency components, the modulation transfer function is 0.5 or more at a spatial frequency of 0.5 cycles/mm and 0.5 at a spatial frequency of 5 cycles/mm.
9. The radiation image processing method according to any one of claims 1 to 8, characterized in that a smoothing process is performed such that the smoothing ratio is 5 or less. 10) An excitation light source for scanning a stimulable phosphor to stimulate the radiation image stored and recorded therein, a photodetector for detecting this emission and converting it into an electrical signal, and this photodetection. The output of the device is divided into two sides parallel to the main scanning direction of the scanning,
A circuit for obtaining an unsharp mask signal Sus corresponding to an extremely low spatial frequency by simple averaging over a rectangular range surrounded by two sides parallel to the sub-scanning direction, the unsharp mask signal Sus, and the photodetector. From the original image signal Sorg which is the output of , and the emphasis coefficient β, the reproduced image signal S
A radiation image processing apparatus comprising an arithmetic device that performs an arithmetic operation expressed by the arithmetic expression S'=Sorg+β(Sorg-Sus). 11) Claim 10, characterized in that the arithmetic unit is equipped with an emphasis coefficient variable means that increases or decreases the emphasis coefficient β according to the magnitude of the image signal Sorg or the mask signal Sus. radiographic image processing device.